Something fundamental is shifting in global finance. It is not a single innovation or a regulatory breakthrough in isolation, it is the simultaneous convergence of three forces that individually reshaped industries, and together are rewriting the infrastructure of money itself. Open banking has democratized access to financial data. Artificial intelligence has made that data actionable at a speed and scale no human analyst can match. Blockchain has created programmable, borderless settlement rails that operate 24 hours a day, seven days a week, without institutional intermediaries. When these three forces converge, as they are converging now in 2026, the result is an architectural transformation.
This report presents NovaVista Capital’s research across this convergence. We examine the competitive landscape of open banking in the United States and Europe, the deployed applications of artificial intelligence across the financial data layer, the blockchain and stablecoin infrastructure being built on top of it, and the profound risk landscape that investors must navigate. We close with our perspectives on where this is all heading and how fast.
The Open Banking Market: A $150 Billion Opportunity Taking Shape
Open banking began as a regulatory mandate, European PSD2 in 2018, the UK’s Open Banking Implementation Entity, and the United States’ slower, market-led equivalent, and has matured into a genuine commercial infrastructure. The global open banking market is estimated at $30 to $35 billion in 2025 and is projected to reach $125 to $150 billion by 2033, growing at a compounded annual rate of 16 to 19 percent. Account-to-account payment volumes, the most commercially significant segment, are expected to grow from $3.2 trillion in 2025 to $5.7 trillion by 2029, a 230 percent increase in four years.
The product architecture of open banking operates across five layers. At the foundation is data aggregation: bank account linking, transaction categorization, and identity verification. Above that sits payment initiation, where account-to-account transfers replace card rails for checkout, subscriptions, and business payments. The third layer is credit decisioning, where cash-flow data replaces or augments credit bureau scores. Identity and fraud prevention operates across all layers. And at the top, embedded finance and banking-as-a-service products allow any company to offer financial services through white-label infrastructure.
The competitive landscape in the United States is led by four players. Plaid sits at the top with an estimated $800 million in revenues, 8,000 application integrations, and connections to 12,000 financial institutions. Its June 2025 partnership with Experian — embedding cash-flow signals directly into consumer credit reports — represents the most commercially significant open banking credit event in US history. Finicity, owned by Mastercard, holds the dominant position in mortgage income verification with direct integrations into Fannie Mae and Freddie Mac’s automated underwriting systems. MX Technologies serves the bank-facing segment with 2,000 financial institution clients. Yodlee, owned by Envestnet and now under Bain Capital, holds legacy market share but is losing developer momentum to Plaid.
In Europe, the landscape is defined by a different dynamic: Visa and Mastercard have both made decisive acquisitions. Visa acquired Tink in 2022 for €1.8 billion, giving it the largest pan-European open banking network across 3,400 banks in 30 countries. Tink is now embedded in Visa’s multi-chain stablecoin settlement infrastructure, making it the most blockchain-converged open banking asset in the world. Trustly, with approximately $350 million in annual revenue and 8,100 merchants, holds the strongest commercial position in account-to-account e-commerce payments. TrueLayer pioneered Variable Recurring Payments in the United Kingdom, the most technically advanced open banking payment product currently live anywhere, and is expanding into Europe. GoCardless has built a dominant position in B2B recurring payments across 90,000 businesses.
The single most important regulatory variable in the United States is the fate of CFPB Section 1033. The rule, finalized in October 2024 and mandating free consumer data access via standardized APIs, was effectively suspended by the Trump administration in early 2025. An interim final rule is expected in the first half of 2026, and its most consequential provision is whether banks will be permitted to charge third-party aggregators for data access. JPMorgan has already signed fee-for-data agreements with many players. If the interim rule codifies broad bank fee-for-data rights, the unit economics of pure-play aggregators break, and the market consolidates around players with Mastercard or Visa balance sheets behind them. In Europe, PSD3 is moving toward harmonization, the EU Instant Payments Regulation is live and mandating 24/7 euro transfers, and MiCA has provided the stablecoin framework that enables the next chapter of open banking’s evolution.
Artificial Intelligence: From Tool to Infrastructure
McKinsey estimates that generative AI could contribute between $200 billion and $340 billion annually to the banking sector. Within open banking specifically, AI has moved from analytical enhancement to operational infrastructure — deployed in production at the largest financial institutions in the world, delivering measurable outcomes that older analytical methods simply cannot match.
Fraud prevention is the most mature application. Mastercard’s Decision Intelligence Pro platform assesses transaction risk across more than one trillion data points, processing in under 300 milliseconds. Its Consumer Fraud Risk product, now live at nine UK banks including Lloyds, NatWest, Monzo, and TSB, identifies authorized push payment scams, where consumers are deceived into authorizing fraudulent transfers, in real time, delivering a 60 percent improvement in mule account identification (a.k.a. account takeover). HSBC reduced false positive rates by 60 percent using deep learning models trained on its transaction history. JPMorgan attributes $1.5 billion in annual savings to AI fraud prevention. The US Treasury’s AI systems prevented or recovered $4 billion in fraudulent payments in fiscal year 2024 alone. The performance improvement over legacy rule-based systems is not incremental — fraud detection latency has fallen from hours to 200 milliseconds, and accuracy rates have risen from 30 to 70 percent in rule-based systems to 90 to 99 percent in deployed AI models.
AI-powered credit underwriting is the second major deployment domain, and its implications are more transformative. Forty-five million Americans are credit-invisible, they have no meaningful credit file at the three major bureaus. Another 100 million have thin files that result in systematic underpricing or denial of credit. Open banking data, transaction histories, income patterns, spending behaviors — provides a rich alternative signal. Upstart’s AI credit model, using more than 1,600 data variables including open banking transaction data, delivers 40 percent higher approval rates than traditional FICO-based underwriting for the same default outcome. Plaid’s June 2025 LendScore product, embedded directly into Experian Consumer Reports, is the first time cash-flow data has appeared alongside bureau data in mainstream credit decisioning. Finicity’s AI income verification is embedded in both Fannie Mae and Freddie Mac’s automated systems, compressing mortgage processing from weeks to hours for qualified borrowers.
Hyper-personalization is the third domain, though its commercial execution has been more uneven than its strategic promise. Bank of America’s Erica virtual assistant has logged more than two billion cumulative interactions, delivering proactive spending analysis, fraud alerts, and financial coaching. NatWest’s AI personalization engine produced a fivefold increase in customer clicks on personalized product offers by detecting life-event signals, a sudden pattern of home-improvement spending, for instance, in transaction data. The commercial opportunity is real: McKinsey estimates personalization can lift banking revenue by 10 to 15 percent and improve customer satisfaction by 20 to 30 percent.
The most consequential emerging domain is agentic AI, systems that autonomously execute multi-step financial tasks without requiring human instruction at the transaction level. Plaid launched its Model Context Protocol server in 2025, enabling AI agents to access real-time financial data via Anthropic’s and OpenAI’s APIs. Its internal deployments have already demonstrated the commercial viability: an AI Annotator that automates transaction labeling and a Fix My Connection agent that autonomously repaired account connectivity issues, enabling two million previously blocked logins and reducing fix time by 90 percent. The agentic layer is where open banking data, AI intelligence, and blockchain execution rails converge, and it is the most disruptive product category currently being built.
Eight unexplored frontiers remain largely unbuilt. The Financial Digital Twin, a persistent AI model of a consumer’s complete financial life that simulates future states before decisions are made, represents a $10 to $20 billion addressable market. Behavioral Credit Velocity scoring, which uses the rate of change in spending patterns as a leading default indicator, could unlock $50 to $80 billion in more accurately priced credit. Agentic Negotiation Bots that autonomously renegotiate subscriptions, challenge bank fees, and switch service providers on behalf of consumers represent a $5 to $10 billion revenue opportunity. A Federated Credit Consortium, where competing banks train a shared AI credit model using privacy-preserving federated learning without ever sharing customer data, could unlock $100 billion or more in collective credit insight no single institution can observe alone.
Blockchain and Stablecoins: Settling the New Payment Architecture
The scale of stablecoin adoption has crossed a threshold that demands direct comparison with incumbent networks. In full-year 2025, stablecoin on-chain volume reached $33 to $35 trillion — against Visa’s $14 trillion in payments volume and Mastercard’s $10.6 trillion in gross dollar volume for their respective fiscal years, per official earnings filings. On gross transaction flow, stablecoins processed more than Visa and Mastercard combined. That headline, however, requires a critical qualification: when Artemis Analytics stripped out DeFi loops, arbitrage, and internal rebalancing, real-economy stablecoin payment volume in 2025 was approximately $390 billion — roughly 2.4 percent of Visa’s figure. Both numbers belong in the same sentence, because together they tell the complete story: the rails already operate at a scale exceeding the world’s largest card network; what remains is the migration of real payment utility onto them. That migration is already underway. Stablecoin B2B payments grew from under $100 million monthly in early 2023 to over $6 billion by mid-2025 — a 60x expansion in two years. In January 2026, monthly stablecoin on-chain volume reached $10 trillion, versus an estimated $1.2 trillion for Visa in the same month. Meanwhile, total stablecoin supply stood at $266 to $318 billion as of January 2026, up from $208 billion in early 2025, with projections from Bernstein pointing to $2.8 trillion in circulation by 2028. These are not projections about a hypothetical future — they describe a market already operating at institutional scale, now anchored by the federal legal clarity of the GENIUS Act signed on July 18, 2025.
The GENIUS Act established the first federal framework for payment stablecoins. Its core provisions – mandatory 1:1 reserves in cash or short-term Treasuries, Bank Secrecy Act AML/KYC obligations, prohibition on rehypothecation, and priority claims for holders in insolvency — do three things simultaneously: they protect consumers, they subject stablecoin issuers to monetary system oversight, and they convert stablecoin demand into a structural buyer base for US Treasury debt. By mandating Treasury-backed reserves, the legislation effectively makes stablecoin growth a fiscal policy tool. The strategic dimension of this is not widely appreciated: the GENIUS Act is as much a debt financing mechanism as a financial consumer protection law.
The blockchain infrastructure underlying this market is not monolithic. Our analysis identifies a structural bifurcation that investors must understand. Ethereum has become the institutional settlement layer, the Wall Street blockchain. JPMorgan’s Kinexys platform processes more than $1 trillion in annualized volume on Ethereum’s Base Layer 2. The UBS tokenized money market fund allows institutional investors to subscribe and redeem around the clock rather than during banking hours. Société Générale’s Forge issued the first MiCA-compliant euro stablecoin on Ethereum. Visa settles USDC payments across Ethereum, Solana, and Avalanche. Ethereum’s base-layer congestion and gas costs limit its utility for high-frequency consumer payments, but its depth of liquidity, institutional trust, and developer ecosystem are structurally irreplaceable for large-value settlement.
Solana is winning the consumer and merchant payment layer. Its transaction costs of $0.0005, settlement times of two to three seconds, and architecture capable of processing more than one million transactions per second with the Firedancer upgrade position it as the dominant infrastructure for retail stablecoin payments. PayPal moved its PYUSD stablecoin from Ethereum to Solana specifically for speed and compliance-in-a-box token capabilities. Visa uses Solana for USDC settlement. Mastercard’s partnership with MoonPay connects 3.5 billion Mastercard cardholders to Solana wallet infrastructure. Stablecoin supply on Solana grew 567 percent in 2025, reaching $12 billion. Avalanche has carved out a distinct role in enterprise compliance infrastructure — its subnet architecture allows financial institutions to run application-specific blockchains with custom validator sets and compliance rules embedded at the protocol level, giving regulated entities the benefits of public blockchain infrastructure without sacrificing regulatory control. Sui’s object-centric architecture is emerging as the most technically elegant solution for programmable consumer finance, including smart-contract-based recurring payment mandates. The market speculates that Meta is in conversations with SUI for launching its second endeavor into the crypto industry.
The convergence thesis rests on a recognition that open banking and stablecoin rails are not competing but complementary. Open banking provides real identity, verified income, behavioral history, regulatory compliance, and consumer trust, in other words, everything that blockchain-native finance lacks. Blockchain provides programmable settlement, 24/7 availability, cross-border reach, and self-enforcing rules, that means, everything that open banking cannot deliver. The platform that bridges both captures the largest commercial opportunity. Visa is building this with Tink and USDC. Mastercard is building it with Finicity and JPM Coin. The five convergence use cases: AI-underwritten on-chain lending; autonomous treasury management; smart-contract fraud blocking; programmable compliance; and the cross-border financial passport; represent addressable markets that did not exist five years ago.
The self-custodial wallet API is among the most underappreciated opportunities in this entire landscape. Open banking platforms today are entirely blind to the $400 to $500 billion in liquid assets held by 15 million or more active DeFi and crypto-native consumers. The full scale of this invisible population is measurable: River Financial’s August 2025 ownership study — using public filings, custodial address tagging, and on-chain research — estimates that individuals collectively hold approximately 13.83 million BTC, representing 65.9% of all circulating supply, across self-custodied wallets and exchange accounts; Glassnode, Chainalysis, and Crypto.com’s 2026 Market Report place total Bitcoin ownership at 480 to 500 million people globally, with 59% of all crypto wallet users preferring non-custodial self-custody over custodial solutions. A consumer earning $8,000 per month in stablecoin yield and holding $150,000 in on-chain assets looks to today’s open banking credit models like a person with no income and minimal assets. They are underwritten as sub-prime when they are economically prime. The first open banking platform to build a production-grade, multi-chain self-custodial wallet API, that is permissioned, AI-enriched, and verified, will unlock $70 billion in incremental addressable market across credit repricing, mortgage eligibility for stablecoin income earners, cross-ecosystem fraud prevention, and holistic consumer financial management. Plaid is the most naturally positioned to build this. Finicity holds the mortgage channel advantage. Neither has announced it yet.
The Risk Landscape: What the Bull Case Gets Wrong
Every structural transformation of this magnitude carries risks proportional to its ambition. Our comprehensive risk analysis identifies four quadrants that investors must price accurately, because the most optimistic scenario, a $650 billion total addressable market by 2030, prices almost none of them.
Smart contract security is the most critical risk in the entire convergence thesis, and it is the most systematically underweighted. By July 2025, stolen funds from cryptocurrency exploits had already reached $2.17 billion for the year, matching the entirety of 2024 with half the year remaining. Bridge exploits alone account for 40 percent of all Web3 security incidents and $2.8 billion in cumulative losses. Any convergence product that moves value between banking rails and blockchain settlement requires a bridge, the most dangerous component in the stack. Bank-grade security in smart contract infrastructure is realistically a 2030 to 2032 development, not a 2027 one. The institutional adoption timeline for DeFi-integrated open banking products is being consistently modeled too aggressively.
Big Tech represents the second critical risk. Amazon, Apple, and Google have deeper capital bases, more intimate customer relationships, and more comprehensive transaction data than any open banking pure-play. Amazon’s merchant cash advance and embedded finance products already underwrite SMB credit more accurately than any aggregator, using first-party transaction data that requires no consumer consent. Google’s three billion Android users represent a distribution advantage that no fintech company can replicate organically. The base case for pure-play open banking aggregators must account for the possibility that Big Tech becomes a competitor.
The US regulatory vacuum is a structural problem, not a temporary one. The CFPB swings between aggressive consumer protection and near-dormancy on a four-year election cycle. CFPB 1033 is the fourth iteration of a rule that has been finalized, challenged, suspended, revised, and re-proposed over nearly a decade. The CLARITY Act’s Senate stall on stablecoin yield treatment is preventing non-bank corporate stablecoin issuance, DeFi integration investment, and self-custodial wallet API development simultaneously. Building compliance architecture for a legal framework that might change in 18 months is economically irrational. The US convergence market is 40 to 50 percent smaller than the bull case assumes until regulatory frameworks achieve sufficient stability to support ten-year infrastructure investment cycles.
Consumer trust is the fourth underweighted risk. Fifty-seven percent of consumers remain unfamiliar with open banking as of 2025. The credit card ecosystem took 50 years to build consumer protection frameworks, chargebacks, fraud liability limits, Reg E, that feel psychologically real and contractually enforceable. Blockchain’s irreversibility is not merely a technical inconvenience (others say, a feature); it is a qualitatively different risk profile that the non-crypto-native demographic, approximately 200 million US adults, will not accept without legislative consumer protection equivalents that do not yet exist. AI model risk compounds this: credit models trained on 2019 to 2021 data failed badly when post-COVID inflation and rate environments diverged from training conditions, as Upstart and Oportun both demonstrated through 2022 and 2023. Autonomous agents executing irreversible blockchain transactions based on AI models that drift during economic dislocations represent a category of systemic risk with no historical precedent.
The probability-weighted addressable market — accounting for trust deficits, smart contract security timelines, regulatory uncertainty, and Big Tech competitive risk — is approximately $233 billion by 2030, not $650 billion. This is still seven times today’s open banking revenue base. It is still a compelling investment thesis. But it is a fundamentally different thesis than the one implied by the most bullish projections, and it demands a fundamentally different portfolio construction to express it.
Final Thoughts: NovaVista Capital’s Perspectives for the Near Future
We are standing at a genuine inflection point, not the kind that gets declared every year by conference panels, but the kind that historians identify in retrospect as the moment when the architecture changed. The convergence of open banking data infrastructure, AI intelligence, and programmable blockchain settlement is a re-platforming of the financial system on foundations that are faster, cheaper, more transparent, more programmable, and more globally accessible than anything that preceded them.
In the next 12 to 18 months, the resolution of the US regulatory stack is the single most important variable. CLARITY Act White House mediation on stablecoin yield treatment, the CFPB 1033 interim rule, and Circle’s IPO performance will collectively define the institutional adoption trajectory for the next five years. Our base case is a compromise on stablecoin yield — activity-linked rewards that accelerate consumer adoption without triggering catastrophic deposit flight — and a fee-capped 1033 interim rule that preserves aggregator margins while providing banks a modest commercial return on data infrastructure investment. In this scenario, the stablecoin market reaches $400 to $600 billion by end-2026, and the first major US bank publicly reports stablecoin operations as a core business line.
Looking out two to four years, structural market embedding replaces regulatory experimentation. Stablecoin-based settlement becomes standard practice for corporate cross-border payments in at least 30 percent of Fortune 500 treasury operations by 2028, driven by the 98 percent settlement time reductions and cost advantages that JPMorgan, Siemens, and Maersk have already validated in production. Solana’s Firedancer upgrade consolidates its position as the primary consumer and merchant payment rail, with on-chain stablecoin supply exceeding $100 billion. The self-custodial wallet API category sees its first serious commercial deployment in 2027 or 2028, unlocking mortgage and credit eligibility for 15 million or more crypto-active consumers currently mis-scored by traditional models. This, more than any other product, represents the most asymmetric near-term opportunity in the open banking space, it is a $70 billion market with no serious competitor building for it today.
At the five to ten-year horizon, the question is whether the convergence creates Autonomous Financial Infrastructure; AI systems that monitor real-time financial health, forecast cash needs, optimize across payment rails, and execute transactions without requiring human instruction at the individual transaction level. The components of this system are already deployed in production: Plaid’s agentic architecture, JPMorgan’s programmable treasury payments, Mastercard’s real-time AI fraud decisioning. What does not yet exist is their integration into a coherent consumer and business product. When it arrives, and we believe the 2028 to 2030 window is realistic for the first credible version, it will be the most commercially significant financial product since the credit card.
The questions in 2026 are no longer whether open banking will scale, whether stablecoins will achieve institutional adoption, or whether AI will transform financial services. Those outcomes are already occurring. The questions that matter for investors are who captures the value in a converged market, how quickly the risk landscape resolves, and which platforms are positioned at the intersection of all three technology layers rather than just one. The divergence between the ceiling and the floor of this opportunity is, itself, the most important signal the market is offering. The ceiling is real. So is the floor. Navigating between them is the work.
Disclaimer: Research materials from NovaVista Capital are for informational purposes only and do not constitute investment advice or an offer to buy or sell securities. Information may be based on third-party sources and is not guaranteed for completeness or accuracy. All research materials are the intellectual property of NovaVista Capital and may not be reproduced, redistributed, or used for commercial purposes without prior written consent.
